When Ocado was founded, its backers realised from the start that off-the-shelf software alone – whether to run its website or its warehouses – just wouldn’t cut the mustard.
At least, not for the grand ambitions that they harboured.
That belief, though, has taken the company into cutting-edge research into robotics, artificial systems, machine learning, simulation and more, as it bids to perfect its systems – and, potentially, to turn the fruits of this research into a platform that it can use to expand or offer as products or services in their own right.
“We are a technology company, but also a logistics company. We are a retailer and, in the future, as part of being a technology company, we will also be a platform company because that’s effectively what we are building,” says Ocado’s technology director, Paul Clarke.
In other words, there’s much more to Ocado than just vans and warehouses, and it isn’t just implementing technology – it’s inventing it.
“We are building an end-to-end platform that starts with the customer placing their order on a web shop or mobile and ends in their order arriving in a one-hour delivery slot at the kitchen table,” he says.
“Another way of looking at it is that you wouldn’t invest the amount that we have invested to build everything, with Ocado owning all of the intellectual property and almost every single line of source code, if you just wanted to do what we are doing now,” he adds.
Hence, the platform that Ocado has been perfecting for the past 12 years can be used to order, pick and deliver any product, not just groceries, and could conceivably be used anywhere in the world, by either Ocado or a licensee. That encompasses the software that runs two of the largest automated warehouses in the world, the robotics increasingly used in those warehouses and the systems that can time orders from website to delivery to within the hour – guaranteed.
That, at least, is the corporate vision.
“The deal we signed with Morrisons [in July 2013] is going to be the first of many. There won’t be other grocery ones in the UK because it’s exclusive, but in other territories and in non-food we can and will do others,” says Clarke.
Technology development at Ocado is split between the mainstream – keeping the existing systems working reliably – research and development, and “10X” advanced or “blue sky” research. “This is conceptual stuff, but it’s not an indulgence. It’s absolutely essential in order for us to maintain the disruptive lead we have over everybody else,” says Clarke.
Perhaps the most eye-opening technology to be developed at Ocado, by teams at both Hatfield in Hertfordshire and in Krakow in Poland, are the “vision” systems that the company intends to apply to robotics.
“Many applications of robotics, in terms of ‘statics’ or b-axis robots, are often quite learnt or repeatable actions, whether that’s stacking trays of lasagne or picking items off of a moving conveyor belt and putting them into an empty box. It does the same thing each time,” says Clarke.
But Ocado is developing robots that can do more.
“What we want to do with robotics... is to be able to pick products from a 40,000 SKU catalogue into a customer ‘tote’ that has got other items in there, having to consider all sorts of rules to do with safety and segregation. For example, what can go with what, what can’t go on top of what and all that kind of thing,” he says.
On top of that, those robots need to work safely alongside ordinary staff – if it were an all-robot environment, it might be easier – and need to consider the impact that the 25 kilometre per hour warehouse conveyor system will have on goods piling up in the customer tote as they are transported at speed from one side of the warehouse to the other.
That is why Ocado has also been working on technology to develop vision systems for the company’s robots.
“They have to also make decisions on the fly. It could be that until it has ‘looked inside’ it doesn’t know what it’s going to do. Or, there are other applications. For example, when goods come
in to our warehouse and it has a pallet of mixed goods and it has to be able to identify one product from another and what needs to be done to them,” says Clarke.
Hence, the vision systems that Ocado has been developing will enable its robots to be able to identify different goods, classify them and deal with them appropriately. This is now on the company technology road-map and could be put into production soon.
It also potentially leads the company into the realm of driverless technology and the possibility that driverless delivery vehicles won’t be far behind.
“The commercial applications of such technology are huge – not only from a cost point of view, but from a safety point of view,” says Clarke.
Indeed, the complicating factor with driverless technology, as with robotics, are the complexities wrought by humans.
“If all of the cars – or most of them – are robots, then it becomes an easier problem than the one you start with, which is ‘robot cars’ having to cope with lots of irrational humans,” says Clarke.
That, though, is one for the future. For the present, the company’s warehouse systems are being perfected all the time, with a number of “change requests” made every day.
“Obviously, some of those will be about fixing defects, but the majority are deploying new functionality at a surprising rate,” says Clarke.
Indeed, while Ocado Technology might sound like a tinkerer’s paradise, Clarke has to remain focused on business needs, even with 10X research. “We run different kinds of projects with different gestation periods. The majority are business-led – I have a budget like anyone else,” says Clarke.
However, he adds: “There are some speculative ventures. Often that’s because individuals have an idea and knock it up on their own and come to us and, if we like it, we turn it into a proper project. Sometimes they take it all the way to near production when they show it to us.
“A lot of what we do is top-down innovation, but driven by the business streams – steering groups that know exactly what we are working on. We keep them informed and when we reach the point where we think it’s ready we may put it in.
“At other times, it’s tied to tight projects that have got tight deadlines, such as taking the Dordon warehouse [in North Warwickshire] live – projects with hard deadlines and a firm date that we have to hit. We run different kinds of projects with different gestation periods,” says Clarke.
Few deadlines were tighter than the one that saw Morrisons decide that it would venture into online shopping and home delivery, not only with Ocado’s technology, but its infrastructure as well – indeed, Morrisons’ online shopping system is all-Ocado right up to the moment it is loaded onto the Morrisons-branded vehicle.
That exclusive deal – it won’t be doing online services for any other UK supermarket chain – gave Clarke and his team just six months to work out how to accommodate the products of another company in its centralised warehouses, from the web front-end all the way through to bakery goods that Morrisons will need to “feed” into Ocado’s warehousing and delivery systems.
Signed in July 2013 with a go-live date in January 2014, Ocado hit the deadline with 11 minutes to spare. Now, having completed the Morrisons rollout at Dordon, Ocado is working on integrating Morrisons with its Hatfield warehouse so that the company can offer home shopping to its customers around London and the South-East.
More significantly, work is also under way to “re-platform” Ocado’s non-warehouse technology for the cloud so that it can be rolled out to others, either elsewhere in the world or in the UK for non-supermarket applications.
Indeed, Clarke sees the potential of cloud computing as the platform enabling the company to expand into other territories without having to establish an expensive data centre there, or offering a more flexible package to potential customers of its technology.
Then, there’s the work on robotics to perfect and helping to turn Ocado’s other advanced research projects into viable business initiatives.
“That is what makes us an exciting place to work as there is always another ‘loop the loop’ on the rollercoaster. But I wouldn’t swap it for a quieter existence,” says Clarke.
Sometimes, the power of the mainframe is the most cost effective answer. Computing's Peter Gothard puts Computing's readers' questions on the future of the mainframe to IBM's Z13 expert Steven Dickens.
This Dummies white paper will help you better understand business process management (BPM)